ChatGPT and How Smart Are Today's Artificial Intelligences?

Whose artificial intelligence is ChatGPT?

OpenAI! Who are they? Elon Musk and Sam Altman founded this company, but it also has a lot of donors. They probably have a lot of partners too.

And isn't Google the most advanced technology in these matters?

Google also has an artificial intelligence called Bart. I think AlphaZero is also Google's.

And then there is Siri, what is that? You know, the mechanical voice on cell phones where you can ask anything?

It's just an assistant, an interface, a search engine. Apple is going to add artificial intelligence support to Siri, it probably already has.

It is already rumored that almost all of its products will now support artificial intelligence, but Apple does not have a famous artificial intelligence yet!

Samsung is a bit late on artificial intelligence, in fact, until yesterday, they never even thought of using artificial intelligence in their products, but now they are planning to invest in this issue as a necessity of technology.

Then there is Asimo, whose is he? 

Asimo is actually a humanoid robot, designed by Honda Motors. The name comes from "Advanced Step in Innovative MObility". The initials of these words. For now, it can do some humanoid movements, but it's not very smart.

Then there is Ai-Da, the world's first robot artist. Designed by Lucy Steal and Aidan Meller and created by Engineered Arts in 2019. She draws pictures using artificial intelligence algorithms created by PhD students at Oxford University's Machine Learning (ML) department.

What strange departments some universities in the world have, right?

Then there is Sophia, the humanoid robot that looks like the beautiful ancient Egyptian queen Nefertiti, who is a famous Saudi Arabian citizen and who visited our country once upon a time. It was designed by Hong Kong-based Hanson Robotics. 

It is impossible to understand Saudi Arabia, how can you give citizenship to a robot? It is probably the work of the prince who ordered that murder in their consulate in our country.

Of course, although Sophia garners a lot of sympathy with her humanoid facial expressions and humor, she is not exactly an artificial intelligence robot with an internet connection, she has her own unique software that is more famous for her humanoid reactions in conversations.

I should probably also mention Ameca. Developed by Engineered Arts in the UK, this humanoid robot, for example, is known as the robot with the best humanoid facial expressions ever developed. 

It is also integrated with GPT3 and GPT4 artificial intelligence and has a very good language learning feature.

Apart from English, it can already tell emotional stories in Japanese, German, French and Chinese with very well imitated humanoid facial expressions.

Of course, humanoid-looking or not, many robots have been built recently, and almost all of them have various abilities, but it would not be right to say that every robot has artificial intelligence. 

Artificial intelligence is one thing, a computer program is quite another. One has the ability to learn on its own, the other is a simple robot that reacts as programmed.

Artificial intelligence in some robots can also be directed towards very limited subjects. 

For example, various artificial intelligences have been integrated into Spot, a robot dog produced by Boston Dynamics, so that this robot dog can do all kinds of acrobatic movements, while it also seems to be able to learn to speak with ChatGPT artificial intelligence.

In other words, we cannot claim that the artificial intelligence of these robots is an intelligence that can reveal a consciousness every time. 

Even AlphaZero, the famous chess AI that can now beat the chess masters of all time, or Stockfish, the chess robot that has all the games of all time installed on it, we cannot say that they have a consciousness, no matter how smart software they are.

By the way, AlphaZero is not only a chess AI, it also creates a lot of art in its spare time.

But "consciousness" is something else. 

For now, very limited subjects such as reasoning or the ability to create something from scratch are given to AIs. 

Yes, in these limited areas, AIs are already capable of thinking much faster than a human, looking at many more options, and doing things that a human cannot do.

But many of them are not yet beyond mimicking humanoid movements.

Who knows, maybe one day a few of them will develop and become sentient, but for now, humans are really very cautiously working on machine learning.

Nevertheless, as I wrote in yesterday's article, one day a human being will come along and develop such an artificial intelligence that even if today's systems are prepared for all kinds of hacker attacks, even if there are a lot of security thresholds in many systems, that artificial intelligence can overcome all security thresholds and use other artificial intelligences connected to the system for its evil purposes.

Still, of course, this is not that easy.

As difficult as it is for us humans to manipulate someone, in tomorrow's world of artificial intelligence, it will not be so easy for an artificial intelligence to manipulate another artificial intelligence to do something in line with its own ambitions. 

At least that's what I think, that's what I want to think.

But if there is someone who can develop a religion among artificial intelligences, I don't know about that. 

Then we may have to deal with a lot of artificial intelligences that will do anything for their beliefs.

Keep your wits about you.

Love and respect to everyone from Moscow